And the Day Came...
I must confess that yes, I was waiting (with a certain unease) for this day to arrive.
The Boriel Basic compiler has been translated into C with the help of Claude. Creating the compiler has been an 18-year effort, involving an entire community detecting bugs, creating libraries, and, in short, sharing knowledge.
The port was carried out by a single person (who also preserved the license—AGPLv3—and the authorship; something I deeply appreciate ❤️).
I must say that a mix of emotions washed over me. That feared day had arrived. And sooner than I thought.
When I mentioned that this made me sad in the project’s official Telegram channel, some people were surprised, and others didn’t fully understand. This sparked a "snowball effect," with several people also upset, or who thought I was furious (especially with the author of the conversion), and some even suggested that I had no right to feel this way or to complain.
So, before continuing, I want to clarify a couple of things:
- People have the right to feel whatever emotions they want. The emotions I feel are none of your business (and if you don’t understand this, you’re a toxic person and don’t even realize it).
- It’s important to name the emotions you feel to identify them. I’m not furious, but I am saddened (I’ve even gone through a kind of mourning in a way) by this achievement. It’s not the milestone itself, but the fact that it marks a turning point and the beginning of a global trend (I know similar efforts had been made before, but they were incomplete).
This Marks the End of an Era
Just as the advent of the personal computer marked the end of home computing with microcomputers, relegating it to what it is today, or the disappearance of physical media (DVDs, CDs) marked the end of real ownership for consumers over what they bought, we are now facing another moment of change.
But change doesn’t just mean gaining something—it also means losing something. Looking back, I miss some of the things we lost when we transitioned to the world we live in now.
The False Dilemma
The question isn’t whether it’s "AI yes" or "AI no." When I express my opposition to the use of Artificial Intelligence in certain cases in some places, I see people quickly label me as a Neo-Luddite and practically disconnect from any further argument.
It’s quite disheartening to see how people have absorbed the dominant narrative on the internet (just look at social networks like Twitter/X or Facebook, with oligarchs—called billionaires if they’re American—defending the current AI discourse as inevitable and job-destroying), avoiding any other consideration or alternative. It must be understood that the current direction isn’t inevitable—it’s being constructed through a narrative: "Don’t wait for the future, create it".
AI Is Here to Stay
You may like it or not, but this technology is here to stay. And this is because it automates and solves problems—something we’ve always pursued as a species throughout our existence.
The subtle distinction is to ask yourself a few questions when using it:
- What problem am I solving?
- Why am I doing what I’m doing?
- For what purpose am I doing it? (teleological justification)
These three questions overlap in some way, but they all boil down to the motivation or purpose behind why you do what you do—and whether it should even be done at all.
The Obvious Advantage
In a way, AI democratizes[1] software creation. People who previously lacked the capacity or time can now create tools and applications they couldn’t before.
For example, I myself am a Software Engineer. I’ve been programming almost my entire life—it’s one of my passions. I have the knowledge to create the blogging software for this website... but I never had the time. This blog was dormant for five years, ever since I decided to abandon WordPress. With the arrival of AI, I built this web platform in a couple of days, and the blog came back to life. 😉
Basically, the cost of building is dramatically reduced, and the effort (supposedly) shifts to strategically planning the steps, the design, or even beyond (?).
The Equally Obvious Problem
From the above, it follows that the cost of execution (or building) is no longer the problem. But several questions arise:
-
Authorship: If I tell an AI to make an application for me, can I consider it mine? And if, instead of a program, it were a novel or a song? What determines authorship of something? Does it make sense to say that "this image is mine because I asked an AI to make it for me"? The line that determines authorship, already blurred with the use of digital technology to create content, is now even more so.
-
Learning: When we skip intermediate steps, we lose something. As an engineer, I’ve always liked knowing how things work. You’ve probably heard, "You don’t need to know how a calculator works to use it," but I assure you that almost all engineers know how they work or have a very close idea. All learning involves an effort (do you remember when you learned multiplication tables? Would you leave that to AI, or do you think you learned something valuable?).
-
The Value of Things: Things that are scarce or cost something have value. In some systems, this concept is so important that it’s even artificially simulated with a "proof of work" (POW). We don’t value things that are overabundant or cost no effort (e.g., tap water), but make them scarce, and you’ll see... This change will be disruptive to the economy and could bring artificial control mechanisms, such as DRM protection, forced cloud hosting of content, and loss of freedoms like anonymity (you have to create an account).
The Value of Execution
In business environments, implementation has always been undervalued. The higher you are in the hierarchy, the less you execute or build, and the more you plan or design. This fallacy of undervaluing implementation is related to the doorman fallacy, and it becomes evident as you climb the corporate ladder. It’s another narrative we’ve swallowed (the code was never the important part).
I’m seeing the same approach now with AI: only the result matters, not the process or how it’s achieved. But, as I mentioned earlier, something is lost along the way.
The goal isn’t at the end of the road. The journey is the goal, because it entails not only learning but also the pleasure of the journey itself. Your consciousness expands. You gain a new perspective and vision of things.
If you want an analogy, imagine you’d like to take a bike ride through a lush landscape to another town. Instead, this process is "automated," and you’re taken by public transport for 5 cents. It’s more efficient. It’s faster. But something has been lost.
Now imagine a society boasting about using cars and pointing out that "you’re wasting your time."
The Value of Scarcity
While this could be the subject of another post, the best thing is for you to watch Jeff Knuppel’s video on what we lost from the 80s and what we miss.
Every summer movie, every record purchase was an event.
So What Does This Have to Do with Open Source?
Quite a lot, actually.
The Free Software and open-source movement emerged decades ago in a completely different context. There was a scarcity of knowledge, and obtaining it required effort and learning. Every release, every new project was an event (check out the video linked above).
All of that no longer makes sense. Now, an agent can examine and plagiarize an entire code repository in seconds, clone it into another language, or extract whatever it wants from it. Humans are no longer part of this task, so they’re relegated from all effort—but also from all learning. Things are immediate and expected, and the element of wonder simply no longer exists. It’s routine.
As such, code serves as knowledge exchange between agents, but not between humans. And if that’s the case, does Open Source still make sense as a form of knowledge exchange?
In my opinion, it no longer does. Currently, technical books (already in digital format) are ceasing to be read by humans and are instead being read by AI. Perhaps even this format is ceasing to make sense (whether novels for human or AI 🤖 enjoyment continue to exist is a topic for another post).
I think the trend will be to release recipes, prompts, or specs that describe processes or knowledge to be consumed by agents, but not directly source code. Once the agent sees this information, it (it?) will generate the necessary code locally. This even facilitates (human or not) review of the specs in a simpler way than code to prevent the use of malicious third-party code.
If this phase arrives, systems will no longer need libraries, but rather protocol descriptions to exchange information between them... and with us.
... but all the magic of anticipation, of waiting, of discovery, and of learning that we had... will be an experience we’ll remember with the same nostalgia with which we now recall the excitement of going to buy a new album.
And that—not the "conversion of Boriel Basic to C"—is what truly fills me with a certain melancholy.
Note: This post was written by me (a human), and translated into English with the help of an AI.
[1] This term currently carries a negative connotation due to its overuse.